Microsoft Research Uses Custom Cloud App To Analyze DNA Data
Experimentation used to take Microsoft scientists hours or days. Now it takes them minutes. This breakthrough is enabled by a custom app that encodes and decodes sequencing files.
In March 2019, researchers from Microsoft and the University of Washington demonstrated the first fully automated system to store and retrieve data in manufactured DNA – a key step in moving the technology out of the research lab and into commercial data centers.
DNA data storage is vitally important for two reasons: (1) it provides a newer, much smaller form of storage than traditional data storage methods which might not be able to keep up with the current amount of data generated and preserved (DNA storage could fit a warehouse-sized data center into a space roughly the size of a few board game dice), and (2) a strand of DNA can last for millions of years, making the stored data last much longer than any other type of data storage.
The key to success in storing data onto DNA is the encoding/decoding process. These involve scientists working on a way to systematically and accurately convert binary digital 1s and 0s into four-part base pairs of DNA (As, Cs, Gs and Ts) so the information can be written into a strand of DNA, and then decoded back to digital. While it sounds simple, Dr. Karin Strauss, Principal Researcher at Microsoft and affiliate professor at UW, says this is in fact one of her most complex projects.
“We’ve had a valuable relationship with GenUI … they have done a great job revamping our Azure-based online application … its new features have simplified our users’ workflows and made our research much easier.”
To aid scientists in this important conversion, Microsoft Research contracted with GenUI to build a custom app that would automate the laborious process of encoding and decoding sequenced files, heretofore performed manually by scientists on their local machines. Using Microsoft’s pre-existing web app as a starting point, GenUI software engineers built the architecture to manage virtual machines that create and destroy on demand to run the encoding and decoding process, all within an intuitive web app. They also replaced shared spreadsheets and brittle folder structures with interfaces and tools to manage the team’s physical pools of DNA and sequencing files. The result? Increased collaboration, faster results and iteration times, and significantly less human error.
One of the key benefits GenUI provided was to systematically and incrementally improve the technical stability of the project. They did this by increasing test coverage from 0-70%, rebuilding the front end of the web app using React and Microsoft UI Fabric, a React library with components styled in Microsoft’s look-and-feel. They also rebuilt the data model in Azure Table Storage and Cosmos DB, all while adding new features and maintaining the core functionality of the app. The research team is now able to maintain their face pace of iteration while simultaneously reducing their technical debt, preventing errors that were threatening to slow them down.
GenUI is a software product engineering firm solving complex technical challenges for technology and software companies – or those who want to become them.